Just to add another data point to the discussion in Rachel’s post…
It seems likely that part of the reason for the decline in men’s LFPR (Labor Force Participation Rate) is the general increase in women’s LFPR over the same time period. Obviously, more women working in the paid labor force means that more husbands will have the option of being supported by their wives. This is not a bad thing in and of itself; it’s only problematic if there’s an increasing trend of households in which women do all the work (paid and unpaid) and men do little or none.
UPDATE: Half Sigma writes:
I respect these guys who are enjoying their leisure instead of working. They haven’t let themselves be brainwashed by conventional middle class values which say that every man has to work otherwise he’s a loser.
As long as the guy isn’t financing his leisure in an abusive or unfair way, I agree.
Pingback: feminist blogs
Don’t you think the dominant factor is more likely to be increasing life expectancy, increasing time in education and longer retirements?
Just in case I’m not the only one who said to themselves “what exactly is the LFPR?” I will save y’all some time in finding out…
Two consistent definitions of the LFPR are here, and here.. The second link also has some male/femal charts, FWIW.
Thanks for the links, Sailorman. [sheepish] I guess I should have explained what the LFPR was in the post… [/sheepish]
Nik, I agree that all three of those factors are virtually certain to be more important reasons for the decline in men’s LFPR.
For those interested:
http://www.scb.se/templates/tableOrChart____27543.asp
Here are some diagrams of worked hours by women (kvinnor) and men (män), 20-64 years old between 1970-2000 in Sweden.
Amp,
I have a slight disagreement with you. I agree with you here, “This is not a bad thing in and of itself; it’s only problematic if there’s an increasing trend of households in which women do all the work (paid and unpaid) and men do little or none.” Unfortunately, my own experience is that there are few cases where this actually is a truly egalitarian work division in practice, but in theory I think there is absolutely nothing to be worried about if the decline in LFP coincides with an increase in unpaid labor. But I have a slight disagreement with you here, “It seems likely that part of the reason for the decline in men’s LFPR (Labor Force Participation Rate) is the general increase in women’s LFPR over the same time period.” Let me explain.
My concern with this is the idea that it is a sort of zero sum game mentality–more jobs for women=fewer jobs for men. On the surface this sounds fine, but it is very important on this issue to think about the profound differences between working class and middle/upper income men. Working class men’s labor force participation has declined mainly because they types of jobs they once occupied are disappearing here in the US. Given the rise of the global economy high paying factory jobs are leaving, and the man featured in the story exemplifies this sort of shift. While high paying(high is relative here–I mean high for someone with a HS diploma) factory jobs are decreasing, lower paying service jobs, which have been disproportionated occupied by women, are actually increasing for the working class. Here is a good Bureau of Labor Statistics chart that details the unemployment rates in various occupations, note how high it is in manufacturing relative to the other industries. I worry that many of the men in this group are averse to taking these lower paying sorts of “pink ghetto” jobs, partly because of the low pay, and partly because these jobs are considered feminine.
From what I know, the loss of “traditionally male dominated jobs” in the middle class is limited. The only major change is that jobs that once excluded women–doctors, attorneys, etc. have become much more gender integrated. However, I think the downward slope of the line you have in this graph would be much less dramatic for middle income men, if the graph was dissected by class.
I have the feeling that some of these working class men are often opting out because they have given up on finding jobs that are becoming really hard to find, and they have this sort of patriarchal I don’t want to go beneath me sort of belief system, which seems to make opting out of the labor force more desirable than taking a low wage job. On the other hand, the middle income men who are opting out probably have a different set of issues. They may not want to take jobs that are “beneath,” but I would suspect that many of them have partners with incomes that can support a family, albeit tenuously.
(I am also certain that this trend is significantly more pronounced among native born Blacks and Latinos.)
Rachel,
I’m surprised that doctors and attorney’s are placed in the middle class (as far as jobs go). In my area physicians and attorneys are the upper class, and the manufacturing jobs (such as my own) are the middle class. You mentioned that those jobs are dissapearing, just as mine might dissapear by the end of the year.
Mendy said, “In my area physicians and attorneys are the upper class, and the manufacturing jobs (such as my own) are the middle class. You mentioned that those jobs are dissapearing, just as mine might dissapear by the end of the year.”
Yes, the vast majority of sociologists would consider professional occupations such as doctors middle class. We usuallly divide the middle into two groups, upper middle class and lower middle class, with jobs like teachers or social workers in the lower middle class and doctors or lawyers in the upper middle class. The upper class would consistent of only the most high paying jobs and quite a few people whose wealth is inhereted. CEOs, CFOs, the highest level managers in corporations, owners of medium and large businesses.
Factory workers, even well paid ones, are generally considered part of the working class.
Generally speaking there are four factors that influence social class–education, occupation, income, and wealth.
Thanks for the clarification Rachel. I was looking at it from the purely practical. In my area, in the deep south, the professions (doctors, lawyers, Ceo’s, CPA’s, etc.) are in the upper class because of their education and their wealth. There are some families that are generally referred to as “old money” and they are the wealthy.
The factory I work at pays execptionally well, (better than teachers, nurses, etc.) and so in our area those employed by this company are considered to be in the middle class, and generally belong to country clubs, drive Escalades, and live in large houses. I saw the end of this era coming and so I bought a cheaper car and haven’t bought a house yet. I’m just not used to seeing 85-90k per year salary considered working class.
Once again thank you for the clarification, now that I understand how the discipline of sociology breaks down class I won’t be talking at cross purposes.
um…
starting the Y-axis of a percentage-over-time plot out at 75% may be useful and necessary in some cases, but it really ought to be noted explicitly. as it stands, that plot is designed to exaggerate and highlight the changes in the dataset; it should be pointed out that, were you to start the plot at the more commonly accepted and understood origin point, it would end up looking far less dramatic.
just my humble opinion. now, off to the library to check out a copy of “how to lie with statistics”…
Nomen,
That’s a good point. I’m sure it is unintentional, but nevertheless it is important.
Point well taken. In my defense, the Bureau of Labor Statistics site I used to generate the chart didn’t give me any control over that aspect of design.
Yes, well we know that there are lies, damned lies, and statistics.
Actually, the tight range does make it possible to see that it fell 5% in the first 20 years, 5% in the next 15 years, and hasn’t fallen close to 5% in the last 20%, which would be harder to see if the y-axis scale were 0-100%.
The period from 1960-1975 seems to have had the greatest rate of decrease. Before and after that, the change seems to be much more gradual. I don’t know anywhere near enough to posit why that was a period of particularly great decrease in labor force participation for men.
Nomen,
That sort of scale expansion is appropriate when you’re looking at up/down changes in a metric that doesn’t fluctuate much, like this. In other words, when the important thing is rather than (or in addition to) an ultimate quantity these scales are handy.
You could show global warming on a kelvin scale, for example, or a F or C scale. But none of those woud do as good a job of allowing you to see a small change in temperature.
I don’t know that I would set the Y-axis to zero, but maybe 50 would be better.
Here is an example of a chart where I think the manipulation of the Y-axis makes a very big difference in how the graph was interpreted. It covers the rate of intermarriage. I use this in my class as an example of a bad graph, note that the y-axis only goes up to 5 %.
“The period from 1960-1975 seems to have had the greatest rate of decrease. Before and after that, the change seems to be much more gradual. I don’t know anywhere near enough to posit why that was a period of particularly great decrease in labor force participation for men.”
I’m thinking the Vietnam War might have had something to do with it (removing large numbers of men from the workforce), but I could be wrong.
Bad link, Rachel.
(BAD link! No clickthrough! Go to your URL right now!)
In fairness to Amp, although he is a despicable and vile twister of the truth who will be the first against the wall when the revolution comes, in this case any despicably vile truth twisting here is the fault of the BLS. It’s their graph generator program, not Amp’s choice of axis scaling.
Robert, It’s fixed. I don’t think Amp is twisting the truth….I just think people should be aware of the Y-axis issue.
Rachel: I’ve seen a lot of “bad graphs” out there, but why do you choose that one as an example? It’s not all that bad to me, and I’m curious. You couldn’t show up to 100% without drastically reducing the scale of the graph to fit the space, or making it 20 times as tall. The former would make it very difficult to read. You would be forced to look at the numbers in order to understan the minuscule difference between the bars. That avoids bias, but it’s not a graph worth making–might as well leave the info in a table.
Of the various “dishonest” ways I could imagine presenting that race information, the displayed graph is probably right near the “honest” end. Hell, I’m more used to seeing graphs which:
1) change scale (logarithmic or, shockingly, literal gaps in the x or y axis)
2) Use different colors/bar widts to imply numbers other than what they show
3) don’t label clearly, causing you to misinterpret
4) show quantized data only (e.g. “Number of people in interracial marriages” without showin the data needed to determine trands (e.g. “US Population”)
5) Show misleading stats like “percent increase” or “percent of mean”
and so on.
Yeah, that graph isn’t so hot because they didn’t include a note about the truncated Y axis.. But email me some time and I’ll try to track you down some really, really, atrocious ones :)
what do you teach? I could talk about bad stats all day :)
Rachel:
Hopefully you know by now that I enjoy your posts and respect your arguments, so this friendly tweak will be seen as funny (as intended) and not obnoxious (not intended).
One of the things which can be pretty misleading is showing multiple graphs with a small change that’s not immediately evident. One major bonus of a graph is that it shouldn’t require a lot of thought to get data. So it’s bad to, say, show two graphs on the same page–both the same size, with the same X axis, and the same subject groups, and the same different numbers on the Y-axes–without being REALLY careful to draw attention to that fact and avoid misreading of the graphs.
Like, say, this page.
hee hee. :D Sorry. But hey–at least I know what you teach :) so you don’t have to answer that part